For many years I've also sought good, solid statistics on things like re-circumcisions, but they're difficult to come by because of no consistent reporting mechanism. Plus, for reasons of shame or privacy, many families do not wish to discuss that the first circumcision didn't go "right".
Williams & Kapila in the UK has been cited for many years as a good guideline of rates of significant complications; Schoen's AAP task force even cited them in their 1989 statement, even though they allegedly "mistyped" the rates as 0.2-1.0% instead of 2.0-10.0% and called it "all complications" instead of "significant complications". Hmmm. Patel in Canada found a much higher rate, 55%, for all levels of surgical complications from infant circumcision, from oozing and bleeding through re-circumcision. Note that Patel refers to phimosis requiring re-circumcision; it may be counterintuitive to many, but phimosis is a common complication of circumcision. The infant body reacts to the wounding by burying the glans even deeper into the skin, which then heals around it pathologically.
My own stab at the US rates, based on many, many years of speaking with doctors and working health fairs and baby fairs (where parents share stories) is that it breaks down something like this:
General complications (requiring a follow-up visit or more): about 25%
Significant complications (requiring a further procedure, such as lysing, or more): about 10%
General re-do's: about 4%
And these figures don't even include meatal stenosis, which I've heard from Peds and FPs can be a high as 40% or more, depending on the degree. Stenosis that requires surgical re-opening may be as high as 5%, but all states of stenosis impact urine flow and potentially reflux.
It's actually rare that I work an event where a parent doesn't come up to me and tell me that their son had to be re-circumcised. It's so common it's frightening, but the medical profession just treats it as part of the deal. Hey MGC ain't neat or pretty.